Outcome | Injury | Severe Injury (time loss >7 days) | Concussion | Severe concussion (time loss >10 days) | ||||

Body checking | No body checking | Body checking | No body checking | Body checking | No body checking | Body checking | No body checking | |

No of players | 608 | 396 | 608 | 396 | 608 | 396 | 608 | 396 |

No of injuries or concussions | 129 | 31 | 78 | 16 | 54 | 17 | 31 | 9 |

Player participation (hours) | 16 162.47 | 8464.91 | 16 162.47 | 8464.91 | 16 162.47 | 8464.91 | 16 162.47 | 8464.91 |

Injury rate, injuries per 1000 player-hours (95% CI)* | 7.98 (4.34 to 11.63) | 3.66 (0.25 to 7.07) | 4.83 (2.02 to 7.63) | 1.89 (0.00 to 4.31) | 3.34 (0.76 to 5.93) | 2.01 (0.00 to 4.78) | 1.92 (0.00 to 4.06) | 1.06 (0.00 to 3.26) |

IRR (unadjusted for covariates, 95% CI) † | 1 (Reference) | 0.46 (0.27 to 0.77) | 1 (Reference) | 0.39 (0.22 to 0.74) | 1 (Reference) | 0.60 (0.31 to 1.19) | 1 (Reference) | 0.55 (0.25 to 1.43) |

Absolute risk reduction associated with removal of checking, injuries per 1000 player-hours | 4.32 (1.80 to 5.83) | 2.94 (1.27 to 3.76) | 1.33 (−0.64 to 2.31) | 0.85 (−0.82 to 1.45) | ||||

No of injuries saved in one season in Alberta‡ | 747 (312 to 1007) | 508 (220 to 650) | 230 (−111 to 399) | 148 (−143 to 250) | ||||

No of injuries saved in one season in Canada§ | 6386 (2667 to 8615) | 4340 (1885 to 5555) | 1970 (−951 to 3416) | 1264 (−1220 to 2137) |

*Crude rates with 95% CIs with a design effect that takes into account clustering (offset by game-hours).

†IRRs and their corresponding 95% CI based on multilevel Poisson regression with one random effect at a team level (offset by game-hours).

‡The population of Bantam players from Alberta in the 2016–2017 season is 7435. The average game-hours was estimated to be 38.75 per Bantam player. The sample proportion used in this study was identified as the lower 60% based on player participation; therefore, the number of players used in the injury reduction calculation is 4461. The number of injuries saved is calculated using Alberta population participation and the absolute risk reduction. The number of injuries is rounded to the nearest integer.

§The population of Bantam players from Canada in the 2016–2017 season is 63 587. The sample proportion used in this study was identified as the lower 60% based on player participation; therefore, the number of players used in the injury reduction calculation is 38 152. The average game-hours was estimated to be 38.75 per Bantam player. The number of injuries saved is calculated using Canada population participation and the absolute risk reduction. The number of injuries is rounded to the nearest integer.

IRRs, incidence rate ratios.